The present invention relates to a sputtering apparatus and method for fabricating durable dielectric thin film optical coatings such as those commonly used in quarter wave stacks for laser mirrors.
In ring laser gyroscopes, one of the principal obstacles to overcome is the phenomena of frequency synchronization or lock-in between the oppositely propagating laser waves within the ring laser cavity. The phenomena of lock-in is fully explained in the test "Laser Applications" edited by Mote Ross, Academic Press, New York, 1971, in the chapter entitled, "Laser Gyroscope," at pages 148-153. The principal cause of lock-in is the phenomena of back scatter which occurs at the reflecting surfaces of the quarter wave stacks in the mirrors in the laser path. Back scatter is caused by anamolies and surface roughness in the reflective surfaces of the various layers of the quarter wave stacks.
Quarter wave stacks and their design are explained in detail in the Military Standardization Handbook entitled "Optical Design," MIL-HDBK-141, Oct. 5, 1962. Briefly, each layer or thin film dielectric coating in a quarter wave stack has a thickness of about one quarter of a wavelength of the light which it is designed to reflect. The number of layers which comprise the quarter wave stack depends on the degree of desired reflectance and the differences in refractive indices of the layers. To increase reflectance, the number of layers and/or the differences in refractive indices may be increased. For mirrors used in ring lasers, the quarter wave stacks generally consist of 17 to 25 quarter wave thin film optical layers deposited on a substrate. Each layer is typically from 500 to 800 Angstroms thick. The layers alternate between a material of high index of refraction and a material of low index of refraction. Typically, the high index material is tantalum pentoxide (Ta.sub.2 O.sub.5) or titanium dioxide (TiO.sub.2) and the low index material is silicon dioxide (SiO.sub.2, i.e., quartz).
In order to minimize back scatter and absorption losses in the quarter wave stack, it is desirable to obtain amorphous coatings which are free of voids and which approach the density and refractive index of the bulk material from which the coatings are obtained. The goal is to get a molecule-by-molecule deposition of the coating and avoid a crystalline structure. Also, it is desired to avoid the formation of suboxides which may result from the lack of sufficient oxygen in the chamber.
Up until now the principal method of fabricating quarter wave stacks for ring laser mirrors has been to use an electron beam evaporation technique. A substrate on which a reflective stack is to be coated is located inside of a vacuum chamber with a sample of the bulk or target material which is to be deposited. An electron beam focussed on the sample material causes localized heating of the material to a point where molecules are evaporated off. These molecules then condense on the other surfaces located in the interior of the vacuum chamber, including the substrate which is being coated.
An electron beam has been used because of its capability of transferring sufficient thermal energy to a localized area of the target material. The kinetic energy of the electrons in the beam are converted to thermal energy when the beam is directed at the evaporate material. Molecules of the target material are heated to the point where molecules or groups of molecules are boiled off. This process of electron beam evaporation as a means of coating is thoroughly explained in the text, "Physical Vapor Deposition," distributed by Airco Temescal, 2850 7th Street, Berkeley, Calif., 1976.
One of the principal problems encountered in the electron beam evaporation technique is to coat layers of the quarter wave stack so that they approach the density of the bulk material from which they come. With this process, molecules of the target material condense on the substrate in such a manner that voids are left between them. The resulting coating is less dense than the bulk, which results in a difference in the layer's index of refraction. Because of the unpredictability of the final density, it is difficult to determine and to control the refractive indices of the stack.
Another problem with the electron beam technique has to do with the electron beam encountering impurities or air pockets in the target material. The high heat concentration results in small explosions which throw out larger chunks of multiple molecules and impurities which condense in the layer. These impurities increase back scatter and absorption in the laser mirror.
With the electron beam evaporation technique, parameters including temperature of the substrate, partial pressure of oxygen in the chamber, rate of deposition, and preparation of the target material are varied in attempts to control and improve the oxidation state, packing density, and degree of amorphousness of the stack. Generally, determining the proper variations and controlling them is very difficult. Typically, temperature of the substrate must be maintained at about 300 degrees Centrigrade in order to get a high density coating which is relatively free of voids and sufficiently amorphous. Until now, the electron beam evaporation technique has been refined to the point where it can consistently produce laser mirrors with losses from absorption and back scatter in the range of 0.1 percent.
Radio frequency (RF) sputtering has been tried in the past as a possible means of depositing thin films for laser mirrors. RF sputtering is explained in Physical Vapor Deposition, supra, at pages 106 to 108. Briefly, the method employes two plates with argon gas between them. On one plate is mounted a substrate to be coated and on the other is the target material. A high frequency, high voltage, a-c field between the plates ionizes the gas atoms causing them to move back and forth striking the target and knocking off molecules which are then deposited on the substrate.
Coatings made in this fashion have tended to be crystaline. Further, the process causes the coating to agglomerate (i.e., have a high surface roughness) and substrate temperature is nearly impossible to control accurately. RF sputtering is presently used for commercial applications, but generally not for specialized thin optical film applications.